The Role of Sprouts in the Restoration of Atlantic Rainforest in Southern Brazil

Rainforest in Southern Brazil

Cecı´lia G. Simo˜es

and Ma´rcia C. M. Marques

Abstract

The relative importance of sexual reproduction (seed) and sprouting as sources for regeneration in Brazilian Atlantic Forest was evaluated in three different successional forest stages: young forest, immature forest, and late-succes- sional forest. Young plants (10–100 cm tall) of tree species were classified into the following categories: (1) seedlings that are nonsprouting—plants that originated through sex- ual reproduction as seeds; (2) stem base sprouting—plants that sprouted at the base of an existing plant; and (3) underground stem sprouting—plants that sprouted from subterranean stems of an existing plant. A total of 1,030 individuals of 48 species were collected. Underground stem sprouting is the rarest form of propagation, with stem base sprouting somewhat more common and possibly associated with recovery of damaged parts. The greatest contribution to regeneration was due to seeds: 92% of the

individuals counted in 67% of the plant species. However, 13 species were ‘‘facultative’’ sprouters as seedlings and sprouters were observed in this group. The three forest ages differed in the proportion of regeneration strategies;

in immature forest, sprouting was more common (15%) than in young (7%) and late-successional (3%; p < 0.05) forest. In these three forest stages, germinating seeds are the major source of new plants; although sprouting as a reproductive strategy is rare, it is related to recovery after damage of an already existing plant and may be due to previous land use history (agriculture) and low soil fer- tility. Restoration using natural regeneration should con- sider these factors to understand seed arrival as seed is the main source of regeneration.

Key words: asexual reproduction, regeneration modes, restinga, seed reproduction, sprouting, tropical rainforest.

Introduction

The Atlantic Forest, the most diverse of any Brazilian for- est system, originally extended along the entire eastern coast of Brazil and into the bordering regions of Paraguay and Argentina. Due to agriculture and urban growth (70% of the Brazilian population is urban), this biome has been reduced to a small fraction of its former size. Today, the Atlantic Forest is considered a global ‘‘hot-spot’’ of biological diversity and is therefore extremely important for conservation (Myers et al. 2000). Although some rem- nants of Atlantic Forest are currently protected in reserves, the largest tracts of forest were totally or par- tially destroyed in the twentieth century. Because these deforested areas are usually abandoned, forest recovery depends on the complex interaction of allochthonous and autochthonous regeneration processes.

In the ecological restoration of tropical forests, planting forest species based on a conceptual framework of ecolog- ical succession is an efficient strategy for establishing the original forest structure and function (Young 2000). Sev- eral initiatives have been successful in the tropics (Parrotta 1991; Ferreti & Britez 2005), but the high cost of these methods impedes their application in forest

rehabilitation (Metzger 2003). If the degree of disturbance was low and if the forest fragments remain, natural regen- eration (by seed, sprouting, or advanced regeneration) is preferred and ecological processes recover after some var- iable time interval (Parrotta et al. 1997; Reis & Kageyama 2003). Understanding the regeneration process is there- fore essential for developing management plans for the restoration and maintenance of Atlantic Forest.

In disturbed tropical forests, recovery of the original forest diversity and species abundance may be initiated through a variety of mechanisms. The seed bank and seed rain, along with sprouting from existing roots or stems of plants that survived the disturbance, may initiate restora- tion (Alvarez-Buylla & Garcia-Barrios 1991; Martı´nez- Ramos & Soto-Castro 1993; Everham & Brokaw 1996;

Dalling & Denslow 1998). The relative importance of these two regeneration mechanisms (seed or sprouting) will determine the successional process and, ultimately, community structure of the mature forest.

Sprouting may result in several individuals (clones) originating from one parent plant (Silvertown 1987).

Although sexual reproduction will provide greater genetic variability (Harper 1977), vegetative propagation through ramets may result in the survival of resistant individuals (Miller & Kauffman 1998). Sprouts have advantages, including an already established root system with a large surface area for water and resource acquisition and high stored energy reserves. These advantages may confer greater chances of plant survival and recovery of damaged

1Laborato´rio de Ecologia Vegetal, Departamento de Botaˆnica, Universidade Federal do Parana´, Caixa Postal 19031 CEP 81531-980 Curitiba, Parana´, Brazil.

2Address correspondence to M. C. M. Marques, email mmarques@ufpr.br

Ó2007 Society for Ecological Restoration International

parts (Sakai et al. 1995; Kennard et al. 2002) and therefore may offset the benefits gained from sexual reproduction, especially when seedling establishment is limited (Sakai et al.

1995).

Although studies have tended to neglect the importance of sprouting in favor of processes associated with sexual reproduction (Bond & Midgley 2001), some studies have examined the importance of sprouting in tropical forests (e.g., Kammesheidt 1998; Miller & Kauffman 1998).

Although we understand that sprouting often occurs in Atlantic Forests (Sa´ 1996; Tabarelli & Mantovani 1999;

Rodrigues et al. 2004), the importance of this method of propagation has never been quantified. The first step to understanding the role that sprouting may play in forest conservation is to examine the relative investment in seed (sexual reproduction) and sprouting (asexual reproduction) in disturbed environments. With these data, we may then examine the ability of species to recover and consequently the regenerative capacity of the forest. Management plans may then be developed for conservation strategies and, based on those data, may recommend natural or artificial propagation to restore forest structure and function.

In this study, we examined the role and importance of sprouting in regeneration in the Atlantic Forest of south- ern Brazil. Specifically, we answer the following questions:

(1) is sprouting common in this forest?; (2) if so, in what proportion of individuals and species does it occur?; (3) do plants have specific regions (aboveground or below- ground level) from which sprouting most commonly occurs?; (4) is sprouting associated with plant size?; and (5) is the relative importance of sprouting constant along successional stages of forest recovery?

Methods

Study Site

The study site is in a region of Atlantic Forest within a pro- tected reserve (Palmito State Forest, total area of approxi- mately 300 ha) in the municipality of Paranagua´ (lat 25°359S, long 48°329W, at sea level), in the state of Parana´, southern Brazil. The climate (Af in Koppen’s classifica- tion) is tropical and superhumid, with neither a pro- nounced dry season nor frosts (Iapar 1978). The 46-year running average temperature is 21°C. Average annual pre- cipitation is 2,218 mm and is always greater than 60 mm even in the driest month (Britez & Marques 2005). Soils (spodzols) are sandy (90–98% sand) and acidic (3.2–3.9 pH; C. Wisniewski, 2006, Universidade Federal do Parana´, personal communication).

The southern coast of Brazil is in the Atlantic Rainfor- est biome, which includes a complex of vegetation types (mainly forest and scrub) on both upland and lowland areas. In lowland areas, where the soil is comprised of marine (Holocene) deposits, the vegetation type is locally called ‘‘restinga’’ (Araujo 1992). In the Palmito reserve, the vegetation type is restinga forest, very similar in flor-

istics and structure to those typical of lowland Atlantic forests that are not over Holocene deposits (Silva et al.

1994). However, the vegetation in the reserve has been modified by 100 years of human occupation. Historically, agricultural practices were rudimentary, with subsistence cultivation based on slash-and-burn agriculture. In the Palmito reserve, forest cutting, planting, and burning oc- curred successively during several years of pineapple and manioc cultivation. As a result of these practices, today, the reserve is a mosaic of different successional stages in- cluding a very small area of old-growth forest. A 0.5-ha area of each of the three forest successional stages was chosen for the study: young forest (21 years old as of 2003), immature forest (34 years), and late-successional forest (59 years). The young forest contains two, more or less, well-defined strata (upper and lower), with a canopy height of approximately 8 m, has relatively low plant species diversity (Shannon–Wiener’s diversity index H9 ¼ 0.95 for trees >5 cm diameter at breast height [dbh]), and is domi- nated by Ilex theazans. The immature forest has three (upper, middle, and lower) strata, with a canopy height of 12 m and greater diversity (H9 ¼ 2.16). The late-succes- sional forest also has three strata, with a canopy height of 16 m and even greater diversity (H9 ¼ 3.29; F. M. Ramos et al., unpublished data). The three forest sites are within close proximity to each other (approximately 1,000 m).

Sampling

In each of the three forest stages, five 2 3 2–m plots were established along a transect (8 m between plots). Due to the irregular distribution of forest in the Palmito reserve (i.e., patches of early- and late-successional forest are very close), a limited area was defined. In each of these plots, all individuals of tree species that were 10–100 cm tall were collected. Our sampling technique was determined by the asymptote of a species area curve, showing that the sample size was sufficient to include all or nearly all species. Each individual was very carefully excavated to gather both aerial and subterranean parts and uniquely labeled. Plants were taken to the laboratory, where they were identified, counted, and diameter at soil level and height measured. Plants were identified by comparison with either adults in the same area or specimens in the herbarium of the Botany Department of the Universidade Federal do Parana´. Species names and authorities were checked in Missouri Botanical Garden tropical database (www.mobot.org).

In the laboratory, plants were carefully examined to determine connections between two or more aerial parts.

When connections existed, indicating sprouting, the form of the connection was noted as aboveground or below- ground and stem or root. When necessary, anatomical slides were prepared for microscopic examination to determine the form of the connection between the plants.

Plants were assumed to have originated from seed in the

absence of evidence of sprouting.

Data Analysis

Plant density and species richness were counted, and Shannon–Wiener’s diversity index (H9) was estimated for each plot in each forest stage (Pielou 1975). Forest stages were compared by the Sorensen Similarity Index (S). To test whether sprouting was associated with plant size in each forest stage, average diameter and height of plants that had sprouted were compared among the three forest stages (analysis of variance [ANOVA], Tukey; Zar 1999).

Sprouting probability (proportion of sprouting plants) and type of sprouting were compared among areas by contin- gency analysis (Zar 1999). All the tests were two tailed and considered significant at p < 0.05.

Results

Forty-eight species and 1,030 individual plants were col- lected (Appendix 1). The species area curve reached an asymptote at 16 m

, which shows that our sampling areas were sufficient for an accurate estimate of the community (Fig. 1). The three forest stages had a similar species composition: young and immature forests—S ¼ 0.73, immature and late-successional forest—S ¼ 0.70, and the greatest difference was between the young and late- successional forest, where S ¼ 0.49. Richness and diversity (H9) varied among the forest stages, with late-successional forest being the most rich and diverse and the young for- est the least (Table 1). The most common species, by for- est stage, were Gomidesia fenzliana in the young forest;

Ocotea pulchella, G. fenzliana, and Myrsine venosa in the immature forest; and O. pulchella and Calophyllum brasi- liense in the late-successional forest (Appendix 1).

Plants arose in three main ways: (1) from seed, non- sprouting seedlings; (2) stem base sprouting, which arises from the stem above the soil level usually after breakage;

and (3) underground stem sprouting, which arises from an underground stem normally parallel to the soil surface.

Of the 48 species, 32 were strictly seedlings, 3 strictly sprouters (stem base or underground stem sprouting), and 13 ‘‘facultative sprouters,’’ in which plants arose through stem base sprouting, underground stem sprouting, or seed (Fig. 2). Of the 16 sprouting species, 8 were only stem base sprouting (Amaioua guianensis, Clethra scabra, Erythroxylum amplifolium, Guapira opposita, Guatteria australis, Myrcia racemosa, and unknown species 7 and 8),

4 only underground stem sprouting species (Alibertia con- color, Blepharocalyx salicifolius, Conomorpha peruviana, and Myrtaceae sp. 1), and 4 included both sprouting types (G. fenzliana, Ilex theazans, O. pulchella, and M. venosa;

Fig. 2). Of the 1,030 individuals, 92.3% showed no evi- dence of sprouting (n ¼ 950), 6.4% showed stem base sprouting (n ¼ 66), and 1.3% showed underground stem sprouting (n ¼ 14).

Average plant height and diameter among the three regeneration types were different (ANOVA; F ¼ 22.97, p < 0.05 and F ¼ 36.02, p < 0.05, respectively). Average height (± SE) of the seedlings (14.4 cm ± 0.57) was lower than that for both types of sprouters (stem base sprouting ¼ 28.9 ± 2.16 and underground stem sprouting ¼ 23.5 ± 4.7;

Tukey, p < 0.05), which were similar (p > 0.10). The dia- meter followed the same pattern, with seedlings (2.13 cm ± 0.14) less than both stem base sprouters (4.4 ± 0.53) and underground stem sprouters (3.7 ± 1.15; Tukey, p < 0.05).

The three forest stages differed in regeneration strat- egy. Seedlings were much more common than both types of sprouting in all the three forest stages (v

¼ 25.1, df ¼ 2, p < 0.05), even after removal from the analysis of G.

fenzliana (the most common species in these three areas), this result was then verified (v

¼ 19.0, df ¼ 2, p < 0.05).

Only 15% of the individuals in the immature forest showed any type of sprouting (average ¼ 7.2 ± 4.6 sprouters per plot), whereas in the young forest (9.0 ± 6.8 sprouters per plot) and late-successional forest (1.4 ± 0.5 sprouters per plot) this percentage was even lower (7 and 3%, respectively; Fig. 3). In comparing only the sprouting plants, the proportion of stem base sprouting and under- ground stem sprouting differed in the three forests stages (v

¼ 17.8, df ¼ 2, p < 0.05). Stem base sprouting occurred more frequently (66%) than underground stem sprouting (34%) in the immature forest and the young forest (97 and 3%, respectively). In the late-successional forest, only eight plants showed stem base sprouting (3% of the total) and no underground stem sprouting was found.

Discussion

A low frequency of sprouting was typical of all the three forest stages, and clearly, more plants arise from seed than sprouting in this restinga forest. Sprouting appears to be occasional and may be associated with recovery after

Table 1.Structural parameters of the regeneration strata (plants1 m tall) of three successional stages (sampling area¼20 m²each) of Atlantic Rainforest in Paranagua´, southern Brazil.

Young Forest (21 Years Old)

Immature Forest (34 Years Old)

Late-Successional Forest (59 Years Old)

Total abundance (n) 514 287 229

Average (± SD) abundance (plants per plot) 102.8 (± 79.4) 57.4 (± 22.8) 45.8 (± 28.2)

Total richness (n) 24 31 30

Average (± SD) richness (plants per plot) 11.2 (± 5.5) 13.2 (± 5.6) 13.0 (± 1.2)

H9 2.18 2.36 2.42

damage, rather than a dominant form of reproduction. In general, in tropical forests, many more species sprout than were observed in this study. Fifty-two percent of species and 85% of individuals were found to sprout in disturbed forests (Everham & Brokaw 1996) compared to 33 and 8%, respectively, found here. Although shrub and tree species of Atlantic forests frequently propagate by sprout- ing (Negrelle 1995; Cirne & Scarano 2001), in this study, we show that the contribution of sprouting to forest regen- eration is less important.

Sprout rarity in the Palmito reserve may be caused by intensive pineapple and manioc cultivation and associated with characteristics typical of restinga forests. Although slash or burn can stimulate tree sprouting, the continued or excessive use of slash-and-burn agriculture may limit

sprout formation (Hartmann et al. 2002). Additionally, the poor soils typical of restinga (approximately 95% sand with low concentrations of phosphorus and potassium;

Britez 2005) may limit carbohydrate storage in the roots and stem and thus inhibit sprouting. This type of forest may favor species that reproduce sexually (Marques &

Oliveira 2005) rather than by sprouting.

Although sprouting is relatively unimportant here in terms of numbers, plants that arose from sprouting are larger in height and diameter because bud formation only occurs following the development of a vascular cambium.

In addition, in the study area, plants greater than 5 cm dbh had different numbers of stems depending on suc- cessional stage (young plants ¼ 3.69 ± 0.48, immature plants ¼ 1.38 ± 0.82, and mature plants ¼ 1.29 ± 1.47; F.

M. Ramos et al., unpublished data). This suggests that growth confers an advantage in branching and sprouting;

so in this study with an upper size class limit, we may have underestimated sprouting. Other studies have shown that plants arising from sprouting are often larger and more resistant (Goto et al. 1996; Miller & Kauffman 1998) and contribute more to total dominance (volume) (Kennard et al. 2002). Conversely, seedlings often suffer greater mor- tality (Goto et al. 1996; Marques & Joly 2000). Thus, although sprouting is numerically less important than ger- mination as a source of plants, sprouted plants may have a growth advantage by initially being larger and having greater reserves, perhaps improving long-term survival.

Stem base sprouting was more common than under- ground stem sprouting and arose from plants that were damaged in some way. Observations in the study area sug- gest that damage is caused by tree falls, mammals in the understorey (including man), pathogens, or herbivores.

Although sprouting of a damaged part may occur, it is an accidental event and hence most likely to be a survival

Figure 1. Species area curves for sampling areas in this study in the Atlantic Rainforest in Paranagua´, southern Brazil. Curves show asymptotes within the number of samples of this study, verifying that most or all species present were sampled.

Figure 2. Percentage of individuals of each plant species with non- sprouting (seedling), stem base sprouting, or underground stem sprouting, in the three successional stages of Atlantic Rainforest in Paranagua´, southern Brazil. * indicates rare species (abundance of5) and ** indicates very common species (abundance of50).

Another 32 recorded species were 100% nonsprouting (regenerated only from seed).

Figure 3. Percentage of nonsprouting (seedling), stem base sprouting, and underground stem sprouting individuals in the three successional stages of Atlantic Rainforest in Paranagua´, southern Brazil. Young forest,n¼514; immature forest,n¼287; and late-successional forest, n¼229 (v²¼25.1,df¼2,p< 0.05).

rather than reproduction strategy (Sakai et al. 1995; Goto et al. 1996; Bond & Midgley 2001).

Aboveground plant formation that arises from under- ground systems is due to the development of buds on roots or stems (Silvertown 1987). Tree species most commonly arise from root buds (Kammesheidt 1999; Hayashi et al.

2001; Rodrigues et al. 2004), however, surprisingly not in any of the species in the forest studied here. On the other hand, sprouting from underground stem buds, which is normally found in herbaceous plants and shrubs (Rizzini

& Heringer 1966; Mallik 1993; Apezzato-da-Glo´ria 2003), occurred in eight tree species in the Palmito reserve.

Four of the 13 species that reproduce both by seed and sprout are in different successional stages (Appendix).

These species invest heavily in sexual and asexual repro- duction, which may be characteristics of pioneer species (Harper 1977). The large investment in reproduction apparently leads to dominance of these species. Tropical species preferably reproduce by seed (Miller & Kauffman 1998), but it is possible that the potential for asexual reproduction is also present and when manifested, may be considered ‘‘facultative resprouting’’ by providing the advantage of vegetative propagation when sexual repro- duction is limited (Goto et al. 1996). Additionally, plants that reproduce in more than one way may persist and thereby stabilize community composition (Eriksson 2000).

In the late-successional forest, fewer plants had sprouts than those in the immature forest or young forest (3 vs. 15 and 7%, respectively). Although sprouting may be more or less constant in early-successional forests with a pre- dominance of secondary species (Kammesheidt 1998, 1999), some species in late-successional or climatic stages may also tend to sprout (Ferna´ndez-Palacios & Are´valo 1998; Kammesheidt 1999; Kennard et al. 2002). In this study, there was a greater frequency of sprouting in the immature and young forests. This is probably related to the exclusivity of some sprouting species in these forests (Blepharocalix salicifolius, Erythroxylum amplifo- lium, Guapira opposita, and Myrtaceae sp. 1) and with the greater total abundance in the early-successional stages than in the late-successional forest.

Sexual reproduction (seed) and sprouting both influence forest composition and consequently capacity for natural restoration. Nevertheless, soil characteristics and land use history also have an important influence on the type of regeneration that takes place. Specifically, in restinga for- ests in coastal Parana´, where a combination of low soil fer- tility and agricultural practices may have limited sprouting, propagation by sprouting is less important than reproduc- tion by seed. Additionally, species richness in the late-suc- cessional stage (30 species with dbh > 5 cm) is much less than that of undisturbed restinga forest (53 species) (Silva et al. 1994), which suggests that natural regeneration is limited due to the lack of sprouting or not enough time has passed to reach the expected number of species.

In Atlantic forests of Brazil, some attempts have been made at artificial restoration by planting (Kageyama et al.

2003; Ferreti & Britez 2005), but natural restoration is much more common, especially in conservation areas (Tabarelli & Mantovani 1999) where financial resources are limited (Dourojeanni & Pa´dua 2001). Thus, future res- toration plans, especially in areas like the restinga (e.g., with poor soils), must consider land use history as well as the raw material for sprouting, such as viable stumps remaining from forest cutting, and the sprouting ability of the species involved, all of which will influence the speed and quality of forest restoration.

Acknowledgments

We are grateful to Instituto Ambiental do Parana´ for per- mission to work in ‘‘Floresta Estadual do Palmito’’

reserve. Thanks to the Conselho Nacional de Desenvolvi- mento Cientı´fico e Tecnolo´gico for partial financial sup- port. Thanks to James Roper, Renato Goldenberg, and Renato Marques, for comments on previous versions of the article, and to John Parrotta, Ludwig Kammesheidt, and two anonymous reviewers for their constructive criti- cism on the final version of the article. Also, gratitude is extended to Sarah Wilkinson for her thorough language editing. Thanks to Marı´lia Borgo for plant species confir- mation and Fernando M. Ramos and Marina Tadei for field assistance.

LITERATURE CITED

Alvarez-Buylla, E. R., and R. Garcia-Barrios. 1991. Seed and forest dynamics: a theoretical framework and an example from the neo- tropics. American Naturalist137:133–154.

Apezzato-da-Glo´ria, B. 2003. Morfologia de Sistemas Subterraˆneos.

ESALQ/USP, Ribeira˜o Preto, Brazil.

Araujo, D. S. D. 1992. Vegetation types of sandy coastal plains of tropical Brazil: a first approximation. Pages 337–381 in U. Seeliger, editor.

Coastal plant communities of Latin America. Academic Press, New York.

Bond, W. J., and J. J. Midgley. 2001. Ecology of sprouting in woody plants: the persistence niche. Tree: Trends in Ecology and Evolution 16:45–51.

Britez, R. M. 2005. Solos. Pages 35–46 in M. C. M. Marques and R. M.

Britez, editors. Histo´ria Natural e Conservacxa˜o da Ilha do Mel. Edi- tora da UFPR, Curitiba, Brazil.

Britez, R. M., and M. C. M. Marques. 2005. Caracterizacxa˜o geral. Pages 13–17 in M. C. M. Marques and R. M. Britez, editors. Histo´ria Natu- ral e Conservacxa˜o da Ilha do Mel. Editora da UFPR, Curitiba, Brazil.

Cirne, P., and F. R. Scarano. 2001. Resprouting and growth dynamic after fire of the clonal shrubAndira legalis (Leguminosae), in a sandy coastal plain in south-eastern Brazil. Journal of Ecology89:351–357.

Dalling, J. W., and J. S. Denslow. 1998. Soil seed bank composition along a forest chronosequence in seasonally moist tropical forest, Panama.

Journal of Vegetation Science9:669–678.

Dourojeanni, M. J., and M. T. J. Pa´dua. 2001. Biodiversidade: a hora deci- siva. Editora da UFPR and Fundacxa˜o O Botica´rio de Protecxa˜o a` Na- tureza, Curitiba, Brazil.

Eriksson, O. 2000. Functional roles of remnant plant populations in commu- nities and ecosystems. Global Ecology and Biogeography9:443–449.

Everham, E. M., and N. V. L. Brokaw. 1996. Forest damage and recovery from catastrophic wind. Botanical Review62:113–185.

Ferna´ndez-Palacios, J. M., and J. R. Are´valo. 1998. Regeneration strate- gies of tree species in the laurel forest of Tenerife (The Canary Island). Plant Ecology137:21–29.

Ferreti, A. R., and R. M. Britez. 2005. A restauracxa˜o da Floresta Atlaˆnti- ca no litoral do Estado do Parana´: os trabalhos da SPVS. Pages 87–102 in A. P. M. Galva˜o and V. Porfı´rio-da-Silva, editors. Restau- racxa˜o florestal: fundamentos e estudos de caso. Embrapa Florestas, Colombo, Brazil.

Goto, Y., T. Yoshitake, M. Okano, and K. Shimada. 1996. Seedling regen- eration and vegetative resprouting after fires inPinus densiflorafor- ests. Vegetatio122:157–165.

Harper, J. L. 1977. Population biology of plants. Academic Press, London, United Kingdom.

Hartmann, T., D. E. Kester, F. T. Davies, and R. Geneve. 2002. Plant propagation: principles and practices. Prentice Hall, Upper Saddle River, New Jersey.

Hayashi, A. H., A. S. Penha, R. R. Rodrigues, and B. Apezzato-da-Glo´ria.

2001. Anatomical studies of shoot bud-forming roots of Brazilian tree species. Australian Journal of Botany49:745–751.

Iapar (Intituto Agronoˆmico do Parana´). 1978. Cartas clima´ticas ba´sicas do Parana´. Instituto Agronoˆmico do Parana´, Curitiba, Brazil.

Kageyama, P. Y., R. E. Oliveira, L. F. D. Moraes, V. L. Engel, and F. B.

Gandara. 2003. Restauracxa˜o ecolo´gica de ecossistemas naturais.

FEPAF, Botucatu, Brazil.

Kammesheidt, L. 1998. The role of tree sprouts in the restoration of stand structure and species diversity in tropical moist forest after slash-and- burn agriculture in Eastern Paraguay. Plant Ecology139:155–165.

Kammesheidt, L. 1999. Forest recovery by root suckers and aboveground sprouts after slash-and-burn agriculture, fire and logging in Paraguay and Venezuela. Journal of Tropical Ecology15:142–157.

Kennard, D. K., K. Gould, F. E. Putz, T. S. Fredericksen, and F. Morales.

2002. Effect of disturbance intensity on regeneration mechanisms in tropical dry forest. Forest Ecology and Management162:197–208.

Mallik, A. U. 1993. Ecology of a forest weed of Newfoundland: vegetative regeneration strategy ofKalmia angustifolia. Canadian Journal of Botany71:161–166.

Marques, M. C. M., and C. A. Joly. 2000. Estrutura e dinaˆmica de uma populacxa˜o deCalophyllum brasilienseCamb. em Floresta higro´fila do sudeste do Brasil. Revista brasileira de Botaˆnica23:107–112.

Marques, M. C. M., and P. E. A. M. Oliveira. 2005. Caracterı´sticas repro- dutivas das espe´cies vegetais da planı´cie costeira. Pages 169–188 in M. C. M. Marques and R. M. Britez, editors. Histo´ria Natural e Conservacxa˜o da Ilha do Mel. Editora da UFPR, Curitiba, Brazil.

Martı´nez-Ramos, M., and A. Soto-Castro. 1993. Seed rain and advanced regeneration in a tropical rain forest. Vegetatio107/108:299–318.

Metzger, J. P. 2003. Como restaurar a conectividade de paisagens frag- mentadas? Pages 49–76 in P. Y. Kageyama, R. E. Oliveira, L. F. D.

Moraes, V. L. Engel, and F. B. Gandara, editors. Restauracxa˜o eco- lo´gica de ecossistemas naturais. FEPAF, Botucatu, Brazil.

Miller, P. M., and J. B. Kauffman. 1998. Seedling and sprout response to slash-and-burn agriculture in a tropical deciduous forest. Biotropica 30:538–546.

Myers, N., R. A. Mittermeier, C. G. Mittermeier, G. A. B. Fonseca, and J. Kent. 2000. Biodiversity hotspots for conservation priorities.

Nature403:853–858.

Negrelle, R. R. B. 1995. Sprouting after uprooting of canopy trees in the Atlantic Rain Forest of Brazil. Biotropica27:448–454.

Parrotta, J. A. 1991. The role of plantation forests in rehabilitating degraded tropical ecosystems. Agriculture, Ecosystems and Envi- ronment00:1–19.

Parrotta, J. A., J. W. Turnbull, and N. Jones. 1997. Catalyzing native for- est regeneration on degraded tropical lands. Forest Ecology and Management99:1–7.

Pielou, E. C. 1975. Ecological diversity. Wiley-Interscience, New York.

Reis, A., and P. Y. Kageyama. 2003. Restauracxa˜o de a´reas degradadas uti- lizando interacxo˜es interespecı´ficas. Pages 91–110 in P. Y. Kageyama, R. E. Oliveira, L. F. D. Moraes, V. L. Engel, and F. B. Gandara, editors. Restauracxa˜o ecolo´gica de ecossistemas naturais. FEPAF, Botucatu, Brazil.

Rizzini, C. T., and E. P. Heringer. 1966. Estudo sobre os sistemas subter- raˆneos difusos de plantas campestres. Anais da Academia Brasileira de Cieˆncias38:85–112.

Rodrigues, R. R., R. B. Torres, L. A. F. Matthes, and A. S. Penha. 2004.

Tree sprouting from root buds in a semideciduous forest affected by fires, Campinas, Southeast Brazil. Brazilian Archives of Biology and Biotechnology47:127–133.

Sa´, C. F. C. 1996. Regeneracxa˜o em a´rea de floresta de Restinga na reserva ecolo´gica estadual de Jacarepia´, Saquarema/RJ: I—Estrato herba´- ceo. Arquivos do Jardim Botaˆnico do Rio de Janeiro34:177–192.

Sakai, A., T. Ohsawa, and M. Ohsawa. 1995. Adaptative significance of sprouting ofEuptelea polyandra, a deciduous tree growing on steep slopes with shallow soil. Journal of Plant Research108:377–386.

Silva, S. M., R. M. Britez, W. S. Souza, and C. A. Joly. 1994. Fitossociolo- gia do componente arbo´reo da Floresta de Restinga da Ilha do Mel, Paranagua´, PR. Pages 47–56 in S. Watanabe, editor. Anais do 3°

Simpo´sio sobre Ecossistemas da Costa Brasileira. ACIESP, Sa˜o Paulo, Brazil.

Silvertown, J. W. 1987. Plant population ecology. Longman Scientific &

Technical, New York.

Tabarelli, M., and W. Mantovani. 1999. Regeneracxa˜o de uma Floresta Tropical Montana apo´s corte e queima (Sa˜o Paulo-Brasil). Revista Brasileira de Biologia59:239–250.

Young, T. P. 2000. Restoration ecology and conservation biology. Biolog- ical Conservation92:73–83.

Appendix. Tree species and their abundance in three successional stages of Atlantic Rainforest in Paranagua´, southern Brazil.

Family Species

Abundance

YF IF LSF

Anacardiaceae Tapirira guianensis Aubl. — — 12

Annonaceae Guatteria australis A. St.-Hil. 10 — 6

Xylopia langsdorfiana St.-Hil. & Tul. — 1 —

Aquifoliaceae Ilex sp. — 1 —

Ilex theazans Mart. 31 6 3

Arecaceae Euterpe edulis Mart. — — 2

Geonoma schottiana Mart. 2 5 2

Celastraceae Maytenus robusta Reissek 2 12 2

Clethraceae Clethra scabra Pers. — — 3

Clusiaceae Calophyllum brasiliense Cambess. 2 6 29

Clusia criuva Cambess. 30 1 1

Erytroxylaceae Erythroxylum amplifolium Baill. 7 1 —

Euphorbiaceae Alchornea triplinervia (Spreng.) Mu¨ll. Arg. 2 1 —

Maprounea brasiliensis A. St.-Hil. 1 — —

Fabaceae Andira fraxinifolia Benth. — 2 3

Machaerium uncinatum Vell. (Benth.) — 1 —

Lauraceae Ocotea aciphylla (Nees) Mez — — 7

O. pulchella (Ness) Mez 131 79 89

Melastomataceae Miconia cabussu Hoehne — 4 —

M. cinerascens Miq. — — 2

Meliaceae Guarea macrophylla Vahl 1 — 5

Myrsinaceae Conomorpha peruviana A. DC. — 10 1

Myrsine venosa A.DC. 69 56 7

Myrtaceae Blepharocalyx salicifolius (Kunth) O. Berg 2 4 —

Calyptranthes rubella (O. Berg) D. Legrand 2 — 1

Gomidesia fenzliana O. Berg 112 57 6

M. multiflora (Lam.) DC. 1 1 —

M. racemosa (O. Berg) Kiaersk. 4 5 13

Myrtaceae sp. 1 — 2 —

Myrtaceae sp. 2 3 3 —

Myrtaceae sp. 3 5 1 —

Nyctaginaceae Guapira opposita (Vell.) Reitz 69 3 —

Podocarpaceae Podocarpus sellowii Klotzsch ex Endl. — — 9

Rosaceae Prunus myrtifolia (L.) Urb. 5 3 —

Rubiaceae Alibertia concolor (Cham.) K. Schum. — 1 3

Amaioua guianensis Aubl. — — 2

Rubiaceae sp. 1 6 —

Sapindaceae Matayba guianensis Aubl. 4 7 12

Theaceae Ternstroemia brasiliensis Cambess. 18 2 —

Unknown Unknown 1 — 2 1

Unknown 2 — 3 —

Unknown 3 — — 1

Unknown 4 — 1 2

Unknown 5 — — 1

Unknown 6 — — 1

Unknown 7 — — 2

Unknown 8 — — 1

IF, immature forest; LSF, late-successional forest; YF, young forest.